} From Hugues.Leroy@irisa.fr
} Date: Thu, 04 Apr 91 09:09:55 +0200
} For audio support, we have xbiff++ with ressources such as:
}
} xbiff*soundBindings:\
} From:.*jbanatre|/usr/site/share/sounds/oh_my_god.au,\
} From:.*leroy|/usr/site/share/sounds/ahoohoo,\
} From:.*Mailer-Daemon|/usr/site/share/sounds/sorrydave,\
} Subject:.*Restau.*|/usr/site/share/sounds/woo_woo_woo_hoo.au
} xbiff*volume:50
}
} so instead of creating new files host/user/face.au, what about the xbiff way?
My own opinions here:
Face.au audio files take much less processing.
With the resource type info, there is yet another file (list) to search
through for every mail message. If this list is long, then there are a lot
of checks needed to determine if you should be playing a sound.
With the face.au file, one stat() is all you need.
Alternatively, the resource approach allows individual customisation per
user which is really nice, and something the face.au approach doesn't.
What do other people think? I like this idea. I'll grab a copy of xbiff++
to see if the copyright notices make the code reusable, and to see if they
have any fast techniques for processing the list to see if there is a sound
file for this user/host.